Monica

#1 neocloud analyst

[Ken Wattana @KenWattana]

If you put Lucy Guo and Brian Chesky on the same podcast you'd have enough energy to power every data center in the world

[Monica @mon161913]

@chloelisbette I can see the future

[Monica @mon161913]

Maintenance capex will get ya

[Monica @mon161913]

FUN doesn’t have the quality of rides / experiences as a Disney / Universal - has massively underinvested in (maintenance!) capex. Also its target demo is getting its dopamine hits from prediction markets / options trading.

Negative macro trends and shitty underlying assets.

[mandy @mandyxyz123]

Personal assistants taking off could be a surprisingly nice tailwind for @SierraPlatform, @DecagonAI, etc. - as more people use AI to bug retailers / service providers for refunds, businesses will want to use AI to fight back.

Right now there's a lot of alpha in having my Instinct spam for refunds, but I wonder how much edge will be left if my Instinct goes up against Sierra. Money that might have gone back into my pocket will go towards compute😂

[my @JeremyRubin]

i'm at the chillis on a friday night. it's a first date, not super important in the eyes of the algorithm because our compatibility score is only an 8.2/10. my date doesn't judge me, she knows how the game works. it's the only place any of us normies seem to be able to get a reservation these days for our ordinary and boring lives. every minute, the staff come out to sing a hooting and hollering happy 21st birthday -- chillis is the only local restaurant that doesn't deboost 21st birthdays.

[Matt Slotnick @matt_slotnick]

the best way to signal status in the Bay Area is to act noticeably stressed about lack of compute

[Monica @mon161913]

@danielmichaelni ppl remember gopro, peloton, etc.

[Monica @mon161913]

Muse is proof that for consumer AI, user experience > the best model

Apple was right on the thesis - they should've owned this, even more so than Google - but was never able to execute

[Monica @mon161913]

Makes sense given OAI / Ant taking <10GWs of the 20-30GWs coming online this year

[Peter Walker @PeterJ_Walker]

Wider context on the share of spend vs the share of tokens across the 3 major American labs

This is aggregate data across OpenRouter and the split between spend and tokens is pretty stark.

The 3 labs still attract ~70% of all spend while only seeing 27% of tokens used

[Monica @mon161913]

@chloelisbette @arakharazian OAI on top 😤

[Monica @mon161913]

@chloelisbette @EugeneNg @pequityresearch Their ARRs roughly track too... it's not like Ant has 6x the ARR as the chart is suggesting

Yes maybe OAI is higher consumer penetration, but enterprise is where the moolah is

[Monica @mon161913]

Community-adjusted EBITDA

[Jukan @jukan05]

“Anthropic’s gross margins are above 80% before accounting for revenue shared with distribution partners, including Amazon, and the cost of training its models.”

Wow.

[Monica @mon161913]

@BackseatVC That’s thousands of cal events

[DaRazor @akramsrazor]

Supreme intelligence so smart...

The reason operators keep A100s racked at all, given that per-MW economics favor Blackwell roughly 2:1 even after the capital charge, is that they're not competing for the same megawatt. A 1kW liquid-cooled NVL72 rack physically can't go into the air-cooled legacy halls where Ampere lives. The A100 tail exists in the shadow of the power constraint — it monetizes stranded, low-density DC capacity that has no alternative use. Which means the 2029 A100 contract isn't evidence about GPU useful life at all; it's evidence about datacenter power scarcity.

[Monica @mon161913]

FUN doesn’t have the quality of rides / experiences as a Disney / Universal - has massively underinvested in (maintenance!) capex. Also its target demo is getting its dopamine hits from prediction markets / options trading.

Negative macro trends and shitty underlying assets.

[A Capital @ACapitalLP]

Why isn’t $FUN a screaming buy here? Josh Kushner just spent $12bn on the Lakers, $LYV is at the highs, and private markets seemingly can’t get enough of scarce experiential assets (the “hedge” for AI).

Meanwhile, $FUN owns an irreplaceable portfolio of regional theme parks and trades like a busted leveraged roll-up. Deleveraging + normalized EBITDA and the equity math gets pretty interesting…

[Monica @mon161913]

Squeeze the other Leopolds outta this market

[Negligible Capital @negligible_cap]

$WDAY is up 30%

Software squeeze superbowl going on right now.

[Monica @mon161913]

Not necessarily worried about terminal value risk but… growth is never gonna reaccelerate under PE… and what’s the exit plan in 5 years?

[Just Another Pod Guy @TMTLongShort]

I’m not involved with the HCM SaaS name that is reportedly engaged in talks with PE and like I said at the start of the year the easy money shorting software has already been made.

That said let me explain why I think it is a terrible bet for private equity to make even though I absolutely believe sunk-cost at an institutional level makes it very likely someone will buy something like this in the coming months.

A. Say you get an asset at a $60b takeout price in exchange for $10b of ongoing revenue that is unlikely to grow but is very very sticky under current buying regime.

And say you can argue you can absolutely annihilate opex taking FCF to $6b a year.

That’s still a decade before you get paid back if you assume terminal value is zero.

And that’s before we talk IRR.

B. So you better be dam sure terminal value won’t be zero.

C. It usually takes 6-12 months for a deal to close. That’s a lifetime in AI model land.

D. Employees are not retarded and know you will gut opex and will jump ship long before deal closes if they are talented.

Good luck keeping product velocity going.

E. Customers now know you are going to be focused on cash flowing them and will adjust roadmaps accordingly.

F. Model capabilities in one year will be good enough to one-shot HCM/payroll with frontier tokens.

G. In three years that same capability will cost 99% less.

H. In three years HR will mostly be run by agents. In three years most of the software purchasing will be done by an in-house agents. There will be repositories where enterprise agents can share software they built with other enterprises. This will all cost less than a single employee at current rate of token deflation.

Could this all be wrong? Sure. But what’s the probability.

[Monica @mon161913]

Profitability != FCFs

[Vinod Khosla @vkhosla]

Profitability is an admission by a company that they have no place to invest that is better than giving the money back to shareholders. You can generate profits or assets for the future. Current situation around data center investments is the same controversy around building assets for the future or keeping cash on hand. OK for shortsighted quarterly Wall Street but not for companies planning great growth 5-10 years out. Higher risk than keeping cash on hand, but substantially more upside. Pick what you want.